The sensitivity of optimally controlled systems to parameter variations is examined in this paper. Two general approaches toward compensating for these variations are employed. The first, for open-loop systems, involves augmenting the performance index with sensitivity terms and minimizing this combined index. The second approach, an adaptive-type controller, involves estimating those parameter variations that have caused an observed state deviation, and adjusting the control policy in response to this measurement. Numerical examples applied to continuous, discrete-time, and discrete-staged systems demonstrate the effectiveness of these approaches in adapting the control policy to parameter variations.